摘要
将感觉检测看作是感觉神经元对其检测属性所进行的抽取与整合，可提出一种基于属性检测与整合的感觉信息处理模型．利用该模型、传统的特征提取理论和空间频率分析器理论有可能得到统一的解释，它是人工神经网络模型的一个属性语义推广并能对某些复杂的神经操作，如跨感觉属性的神经网络检测与整合、知觉意识中的表象匹配、属性补充与修正机制给出一种较自然的数学表达．
Detection and integration of sensation attribute can be regard respectively as an abstraction and conjunction of the attribute detected by sensation neuron. A new kind of mathematical representation of neural information processing, which is based on detection and integration of attribute, is presented in this paper. An unifying mathematical model of both the spatial frequency analysis theory and the feature abstraction theory can be proposed, and it is an semantic extension of artificial neural work; and some complex nerual operations, for exmple, the neural work detection and integration of Multi sensational complex attribute can be repressented by mathematical language.
出处
《广西师范大学学报（自然科学版）》
CAS
1996年第1期1-7,共7页
Journal of Guangxi Normal University:Natural Science Edition
基金
国家自然科学基金
国家攀登计划基金
广西自然科学基金
关键词
属性检测
整合
数学表达
神经网络
感觉检测
sensation
detection of attribute
integration
mathematical representation
